Testing the Toxicofera: Comparative Reptile Transcriptomics Casts Doubt on the Single, Early Evolution of the Reptile Venom System

Testing the Toxicofera: comparative reptile transcriptomics casts doubt on

ANGOR UNIVERSITY the single, early evolution of the reptile venom system Hargreaves, A.D.; Swain, M.T.; Logan, D.W.; Mulley, J.F.

Toxicon

DOI: 10.1016/j.toxicon.2014.10.004

PRIFYSGOL BANGOR / B Published: 17/10/2014

Peer reviewed version

Cyswllt i'r cyhoeddiad / Link to publication

Dyfyniad o'r fersiwn a gyhoeddwyd / Citation for published version (APA): Hargreaves, A. D., Swain, M. T., Logan, D. W., & Mulley, J. F. (2014). Testing the Toxicofera: comparative reptile transcriptomics casts doubt on the single, early evolution of the reptile venom system. Toxicon, 92, 140-156. https://doi.org/10.1016/j.toxicon.2014.10.004

Hawliau Cyffredinol / General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ?

Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

03. Oct. 2021 1 Title page

3 Title:

4 Testing the Toxicofera: comparative transcriptomics casts doubt on the single, early evolution

5 of the reptile venom system

7 Authors:

8 Adam D Hargreaves1, Martin T Swain2, Darren W Logan3 and John F Mulley1*

10 Affiliations:

11 1. School of Biological Sciences, Bangor University, Brambell Building, Deiniol Road,

12 Bangor, Gwynedd, LL57 2UW, United Kingdom

14 2. Institute of Biological, Environmental & Rural Sciences, Aberystwyth University,

15 Penglais, Aberystwyth, Ceredigion, SY23 3DA, United Kingdom

17 3. Wellcome Trust Sanger Institute, Hinxton, Cambridge, CB10 1HH, United Kingdom

19 *To whom correspondence should be addressed: [email protected]

22 Author email addresses:

23 John Mulley - [email protected]

24 Adam Hargreaves – [email protected]

25 Darren Logan - [email protected]

26 Martin Swain - [email protected] 1

27 Abstract

28 The identification of apparently conserved gene complements in the venom and salivary

29 glands of a diverse set of reptiles led to the development of the Toxicofera hypothesis – the

30 single, early evolution of the venom system in reptiles. However, this hypothesis is based

31 largely on relatively small scale EST-based studies of only venom or salivary glands and

32 toxic effects have been assigned to only some putative Toxicoferan toxins in some species.

33 We set out to examine the distribution of these proposed venom toxin transcripts in order to

34 investigate to what extent conservation of gene complements may reflect a bias in previous

35 sampling efforts. Our quantitative transcriptomic analyses of venom and salivary glands and

36 other body tissues in five species of reptile, together with the use of available RNA-Seq

37 datasets for additional species, shows that the majority of genes used to support the

38 establishment and expansion of the Toxicofera are in fact expressed in multiple body tissues

39 and most likely represent general maintenance or “housekeeping” genes. The apparent

40 conservation of gene complements across the Toxicofera therefore reflects an artefact of

41 incomplete tissue sampling. We therefore conclude that venom has evolved multiple times in

42 reptiles.

44 Keywords

45 Snake venom, Toxicofera, Transcriptomics

47 1. Introduction

48 Snake venom is frequently cited as being highly complex or diverse (Li et al., 2005; Wagstaff

49 et al., 2006; Kini and Doley, 2010) and a large number of venom toxin genes and gene

50 families have been identified, predominantly from EST-based studies of gene expression

51 during the re-synthesis of venom in the venom glands following manually-induced emptying

52 (“milking”) of extracted venom (Pahari et al., 2007; Casewell et al., 2009; Siang et al., 2010; 2

53 Rokyta et al., 2011; Rokyta et al., 2012). It has been suggested that many of these gene

54 families originated via the duplication of a gene encoding a non-venom protein expressed

55 elsewhere in the body, followed by recruitment into the venom gland where natural selection

56 could act to increase toxicity. Subsequent additional duplications would then lead to a

57 diversification within gene families, often in a species-specific manner (Fry, 2005; Wong and

58 Belov, 2012; Casewell et al., 2013). However, since whole genome duplication is a rare event

59 in reptiles (Otto and Whitton, 2000), the hypothesis that novelty in venom originates via the

60 duplication of a “body” gene with subsequent recruitment into the venom gland requires that

61 gene duplication is a frequent event in the germline of venomous snakes. An additional

62 prerequisite is that the promoter and enhancer sequences that regulate venom gland-specific

63 expression are relatively simple and easy to evolve. It also suggests a high incidence of

64 neofunctionalisation rather than the more common process of subfunctionalisation (Lynch

65 and Force, 2000; Walsh, 2003; Lynch, 2007; Teshima and Innan, 2008). However, it has

66 recently been shown that in fact snake venom toxins are likely derived from pre-existing

67 salivary proteins that have been restricted to the venom gland rather than body proteins that

68 have been recruited (Hargreaves et al. 2014a).

69 The apparent widespread distribution of genes known to encode venom toxins in snakes in

70 the salivary glands of a diverse set of reptiles, (including both those that had previously been

71 suggested to have secondarily lost venom in favour of constriction or other predation

72 techniques, and those that had previously been considered to have never been venomous), led

73 to the development of the Toxicofera hypothesis – the single, early evolution of venom in

74 reptiles (Vidal and Hedges, 2005; Fry et al., 2006; Fry et al., 2009a; Fry et al., 2012a) (Figure

75 1). Analysis of a wide range of reptiles, including charismatic megafauna such as the

76 Komodo dragon, Varanus komodoensis (Fry et al., 2009b), has shown that the ancestral

77 Toxicoferan venom system comprises at least 16 genes, with additional gene families

78 subsequently recruited in different lineages (Fry et al., 2009a; Fry et al., 2012a; Fry et al.,

79 2013).

80 Although toxic effects have been putatively assigned to some Toxicoferan venom proteins in

81 certain species, the problem remains that their identification as venom components is based

82 largely on their expression in the venom gland during venom synthesis and their apparent

83 relatedness to other, known toxins in phylogenetic trees. It has long been known that all

84 tissues express a basic set of “housekeeping” or maintenance genes (Butte et al., 2002) and it

85 is therefore not surprising that similar genes might be found to be expressed in similar tissues

86 in different species of reptiles, and that these genes might group together in phylogenetic

87 trees. However, the identification of transcripts encoding putative venom toxins in other body

88 tissues would cast doubt on the classification of these Toxicoferan toxins as venom

89 components, as it is unlikely that the same gene could fulfil toxic and non-toxic (pleiotropic)

90 roles without evidence for alternative splicing to produce a toxic variant (as has been

91 suggested for acetylcholinesterase in the banded krait, Bungarus fasciatus (Vonk et al., 2011;

92 Casewell et al., 2013)) or increased expression levels in the venom gland (where toxicity

93 might be dosage dependent). In order to address some of these issues and to test the

94 robustness of the Toxicofera hypothesis, we have carried out a comparative transcriptomic

95 survey of the venom or salivary glands, skin and cloacal scent glands of five species of

96 reptile. Unlike the pancreas and other parts of the digestive system (Strydom, 1973; Kochva,

97 1987), these latter tissues (which include a secretory glandular tissue (the scent gland) and a

98 relatively inert, non-secretory tissue (skin)) have not previously been suggested to be the

99 source of duplicated venom toxin genes and we would therefore only expect to find

100 ubiquitous maintenance or “housekeeping” genes to be commonly expressed across these

101 tissues. We use the general term ‘salivary gland’ for simplicity, to encompass the oral glands

102 of the leopard gecko and rictal glands and Duvernoy’s gland of the royal python, corn snake

103 and rough green snake and do not imply any homology to mammalian salivary glands. 4

104 Our study species included the venomous painted saw-scaled viper (Echis coloratus); the

105 non-venomous corn snake (Pantherophis guttatus) and rough green snake (Opheodrys

106 aestivus) and a member of one of the more basal extant snake lineages, the royal python

107 (Python regius). As members of the Toxicofera sensu Fry et al. (Fry et al., 2013) we would

108 expect to find the basic Toxicoferan venom genes expressed in the venom or salivary glands

109 of all of these species. In addition, we generated corresponding data for the leopard gecko

110 (Eublepharis macularius), a member of one of the most basal lineages of squamate reptiles

111 that lies outside of the proposed Toxicofera clade (Figure 1). We have also taken advantage

112 of available transcriptomes or RNA-Seq data for corn snake vomeronasal organ

113 (Brykczynska et al., 2013) and brain (Tzika et al., 2011), garter snake (Thamnophis elegans)

114 liver (Schwartz and Bronikowski, 2013) and pooled tissues (brain, gonads, heart, kidney,

115 liver, spleen and blood of males and females (Schwartz et al., 2010)), Eastern diamondback

116 rattlesnake (Crotalus adamanteus) and eastern coral snake (Micrurus fulvius) venom glands

117 (Rokyta et al., 2011; Rokyta et al., 2012; Margres et al., 2013), king cobra (Ophiophagus

118 hannah) venom gland, accessory gland and pooled tissues (heart, lung, spleen, brain, testes,

119 gall bladder, pancreas, small intestine, kidney, liver, eye, tongue and stomach) (Vonk et al.,

120 2013), Burmese python (Python molurus) pooled liver and heart (Castoe et al., 2011), green

121 anole (Anolis carolinensis) pooled tissue (liver, tongue, gallbladder, spleen, heart, kidney and

122 lung), testis and ovary (Eckalbar et al., 2013) and bearded dragon (Pogona vitticeps), Nile

123 crocodile (Crocodylus niloticus) and chicken (Gallus gallus) brains (Tzika et al., 2011), as

124 well as whole genome sequences for the Burmese python and king cobra (Castoe et al., 2013;

125 Vonk et al., 2013).

126 Assembled transcriptomes were searched for genes previously suggested to be venom toxins

127 in Echis coloratus and related species (Wagstaff and Harrison, 2006; Casewell et al., 2009;

128 Wagstaff et al., 2009) as well as those that have been used to support the Toxicofera

129 hypothesis, namely acetylcholinesterase, AVIT peptide (Fry, 2005; Fry et al., 2009a; Vonk et 5

130 al., 2011; Fry et al., 2012a; Casewell et al., 2013), complement c3/cobra venom factor,

131 epididymal secretory protein (Alper and Balavitch, 1976; Fry et al., 2012a), c-type lectins

132 (Morita, 2005; Ogawa et al., 2005), cysteine-rich secretory protein (crisp) (Yamazaki et al.,

133 2003a; Yamazaki and Morita, 2004), crotamine (Rádis-Baptista et al., 2003; Oguiura et al.,

134 2005), cystatin (Ritonja et al., 1987; Richards et al., 2011), dipeptidylpeptidase, lysosomal

135 acid lipase, renin aspartate protease (Wagstaff and Harrison, 2006; Aird, 2008; Casewell et

136 al., 2009; Fry et al., 2012a), hyaluronidase (Tu and Hendon, 1983; Harrison et al., 2007),

137 kallikrein (Komori et al., 1988; Komori and Nikai, 1998), kunitz (Župunski et al., 2003), l-

138 amino-acid oxidase (Suhr and Kim, 1996; Du and Clemetson, 2002), nerve growth factor

139 (Angeletti, 1970; Kostiza and Meier, 1996), phospholipase A2 (Lynch, 2007), phospholipase

140 b (Bernheimer et al., 1987; Chatrath et al., 2011; Rokyta et al., 2011), ribonuclease (Aird,

141 2005), serine protease (Pirkle, 1998; Serrano and Maroun, 2005), snake venom

142 metalloproteinase (Bjarnason and Fox, 1994; Jia et al., 1996), vascular endothelial growth

143 factor (vegf) (Junqueira de Azevedo et al., 2001; Yamazaki et al., 2003b; Fry, 2005; Fry et

144 al., 2006), veficolin (OmPraba et al., 2010), vespryn, waprin (Torres et al., 2003; Pung et al.,

145 2006; Nair et al., 2007; Fry et al., 2012a) and 3-finger toxins (Fry et al., 2003). Transcript

146 abundance estimation values were also calculated to allow the identification of any potential

147 occurrences of pleiotropy (a gene fulfilling a toxic and non-toxic role simultaneously) based

148 upon an elevated expression level in the venom or salivary gland compared to other body

149 tissues. All Transcript abundance values are given in FPKM (Fragments Per Kilobase of exon

150 per Million fragments mapped) and are mean values to account for variation between

151 individual samples (further details given in the methods section).

152 We find that many genes previously claimed to be venom toxins are in fact expressed in

153 multiple tissues (Figure 2) and that transcripts encoding these genes show no evidence of

154 consistently elevated expression level in venom or salivary glands compared to other tissues

155 (Supplemental tables S5-S9). Only two putative venom toxin genes (l-amino acid oxidase b2 6

156 and PLA2 IIA-c) showed evidence of a venom gland-specific splice variant across our

157 multiple tissue data sets. We have also identified several cases of mistaken identity, where

158 non-orthologous genes have been used to claim conserved, ancestral expression and instances

159 of identical sequences being annotated as two distinct genes (see later sections). We propose

160 that the putative ancestral Toxicoferan venom toxin genes do not encode toxic venom

161 components in the majority of species and that the apparent venom gland-specificity of these

162 genes is a side-effect of incomplete tissue sampling. Our analyses show that neither increased

163 expression in the venom gland nor the production of venom-specific splice variants can be

164 used to support continued claims for the toxicity of these genes.

165

166 2. Methods

167 Experimental methods involving animals followed institutional and national guidelines and

168 were approved by the Bangor University Ethical Review Committee.

169 2.1 RNA-Seq

170 Total RNA was extracted from four venom glands taken from four individual specimens of

171 adult Saw-scaled vipers (Echis coloratus) at different time points following venom extraction

172 in order to capture the full diversity of venom genes (16, 24 and 48 hours post-milking).

173 Additionally, total RNA from two scent glands and two skin samples of this species and the

174 salivary, scent glands and skin of two adult corn snakes (Pantherophis guttatus), rough green

175 snakes (Opheodrys aestivus), royal pythons (Python regius) and leopard geckos (Eublepharis

176 macularius) was also extracted using the RNeasy mini kit (Qiagen) with on-column DNase

177 digestion. Only a single corn snake skin sample provided RNA of high enough quality for

178 sequencing. mRNA was prepared for sequencing using the TruSeq RNA sample preparation

179 kit (Illumina) with a selected fragment size of 200-500bp and sequenced using 100bp paired-

180 end reads on the Illumina HiSeq2000 or HiSeq2500 platform.

181 2.2 Quality control, assembly and analysis 7

182 The quality of all raw sequence data was assessed using FastQC (Andrews, 2010) and reads

183 for each tissue and species were pooled and assembled using Trinity (Grabherr et al., 2011)

184 (sequence and assembly metrics are provided in Supplemental tables S1-S3). Putative venom

185 toxin amino acid sequences were aligned using ClustalW (Larkin et al., 2007) and maximum

186 likelihood trees constructed using the Jones-Taylor-Thornton (JTT) model with 500

187 Bootstrap replicates. Transcript abundance estimation was carried out using RSEM (Li and

188 Dewey, 2011) as a downstream analysis of Trinity (version trinityrnaseq_r2012-04-27). Sets

189 of reads were mapped to species-specific reference transcriptome assemblies (Supplementary

190 table S4) to allow comparison between tissues on a per-species basis and all results values

191 shown are in FPKM (Fragments Per Kilobase of exon per Million fragments mapped).

192 Individual and mean FPKM values for each gene per tissue per species are given in

193 Supplementary tables S5-S9. All transcript abundance values given within the text are based

194 on the average transcript abundance per tissue per species to account for variation between

195 individual samples.

196 Transcriptome reads were deposited in the European Nucleotide Archive (ENA) database

197 under accession #ERP001222 and GenBank under the run accession numbers SRR1287707

198 and SRR1287715. Genes used to reconstruct phylogenies are deposited in GenBank under the

199 BioProject accession number PRJNA255316.

200

201

202 3. Results

203 3.1 Genes unlikely to represent toxic components of the Toxicofera

204 Based on our quantitative analysis of their expression pattern across multiple species, we

205 identify the following genes as unlikely to represent toxic venom components in the

206 Toxicofera clade (Vidal and Hedges, 2005). The identification of these genes as non-venom

207 is more parsimonious than alternative explanations such as the reverse recruitment of a 8

208 “venom” gene back to a “body” gene (Casewell et al., 2012), which requires a far greater

209 number of steps (duplication, recruitment, selection for increased toxicity, reverse

210 recruitment) to have occurred in each species, whereas a “body” protein remaining a “body”

211 protein is a zero-step process regardless of the number of species involved. The process of

212 reverse recruitment must also be considered doubtful given the rarity of gene duplication in

213 vertebrates (estimated to be between 1 gene per 100 to 1 gene per 1000 per million years

214 (Cotton and Page, 2005; Lynch and Conery, 2000; Lynch and Conery, 2003)).

215 3.1.1 Acetylcholinesterase

216 We find identical acetylcholinesterase (ache) transcripts in the E. coloratus venom gland and

217 scent gland (which we call transcript 1) and an additional splice variant expressed in skin and

218 scent gland (transcript 2). Whilst the previously known splice variants in banded krait

219 (Bungarus fasciatus) are differentiated by the inclusion of an alternative exon, analysis of the

220 E. coloratus ache genomic sequence (accession number KF114031) reveals that the shorter

221 transcript 2 instead comprises only the first exon of the ache gene, with a TAA stop codon

222 that overlaps the 5’ GT dinucleotide splice site in intron 1. ache transcript 1 is expressed at a

223 low level in the venom gland (6.60 FPKM) and is found in multiple tissues in all study

224 species (Figure 2), as well as corn snake vomeronasal organ and garter snake liver. The

225 shorter transcript 2 is found most often in skin and scent glands (Figure 2, Supplementary

226 figure S1). The low expression level and diverse tissue distribution of transcripts of this gene

227 suggest that acetylcholinesterase does not represent a Toxicoferan venom toxin. Whilst some

228 ACHE activity has been recorded in the oral secretions of a number of colubrid snakes

229 (Mackessy, 2002), experiments with these secretions shows that several hours are needed to

230 achieve complete neuromuscular blockage. It should also be noted that the most frequently

231 cited sources for the generation of a toxic version of ache in banded krait via alternative

232 splicing include statements that ache “does not appear to contribute to the toxicity of the

233 venom” (Cousin et al., 1998), is “not toxic to mice, even at very high doses” (Cousin et al., 9

234 1996a) and is “neither toxic by itself nor acting in a synergistic manner with the toxic

235 components of venom” (Cousin et al., 1996b).

236 3.1.2 AVIT

237 We find only a single transcript encoding an AVIT peptide in our dataset, in the scent gland

238 of the rough green snake (data not shown). The absence of this gene in all of our venom and

239 salivary gland datasets, as well as the venom glands of the king cobra, eastern coral snake and

240 Eastern diamondback rattlesnake, and the limited number of sequences available on Genbank

241 (one species of snake, Dendroaspis polylepis (accession number P25687) and two species of

242 lizard, Varanus varius and Varanus komodoensis (accession numbers AAZ75583 and

243 ABY89668 respectively)) despite extensive sampling, would suggest that it is unlikely to

244 represent a conserved Toxicoferan venom toxin.

245 3.1.3 Complement C3 (“cobra venom factor”)

246 We find identical transcripts encoding complement c3 in all tissues in all species, with the

247 exception of royal python skin (Figures 2 and 3) and we find only a single complement c3

248 gene in the E. coloratus genome (data not shown). These findings, together with the

249 identification of transcripts encoding this gene in the liver, brain, vomeronasal organ and

250 tissue pools of various other reptile species (Figure 3) demonstrate that this gene does not

251 represent a Toxicoferan venom toxin. However, the grouping of additional complement c3

252 genes in the king cobra (Ophiophagus hannah) and monocled cobra (Naja kaouthia) in our

253 phylogenetic tree does support a duplication of this gene somewhere in the Elapid lineage.

254 One of these paralogs may therefore represent a venom toxin in at least some of these more

255 derived species and the slightly elevated expression level of this gene in the venom or

256 salivary gland of some of our study species suggests that complement c3 has been exapted

257 (Gould and Vrba, 1982) to become a venom toxin in the Elapids. It seems likely that the

258 identification of the non-toxic paralog in other species (including veiled chameleon

259 (Chamaeleo calyptratus), spiny-tailed lizard (Uromastyx aegyptia) and Mitchell's water 10

260 monitor (Varanus mitchelli)) has contributed to confusion about the distribution of this

261 “Cobra venom factor” (which should more rightly be called complement c3b), to the point

262 where genes in alligator (Alligator sinensis), turtles (Pelodiscus sinensis) and birds (Columba

263 livia) are now being annotated as venom factors (accession numbers XP 006023407-8, XP

264 006114685, XP 005513793, Figure 3).

265 3.1.4 Cystatin

266 We find two transcripts encoding cystatins expressed in the venom gland of E. coloratus

267 corresponding to cystatin-e/m and f (Supplementary figures S2 and S3). cystatin-e/m was

268 found to be expressed in all tissues from all species used in this study (Figure 2), as well as

269 corn snake vomeronasal organ and brain and garter snake liver and pooled tissues. The

270 transcript encoding cystatin-f (which has not previously been reported to be expressed in a

271 snake venom gland) is also expressed in the scent gland of E. coloratus and in the majority of

272 other tissues of our study species. We find no evidence for a monophyletic clade of

273 Toxicoferan cystatin-derived venom toxins and would agree with Richards et al. (Richards et

274 al., 2011) that low expression level and absence of in vitro toxicity represents a “strong case

275 for snake venom cystatins as essential housekeeping or regulatory proteins, rather than

276 specific prey-targeted toxins…” Indeed, it is unclear why cystatins should be considered to be

277 conserved venom toxins, since even the original discovery of cystatin in the venom of the

278 puff adder (Bitis arietans) states that there is “…no evidence that it is connected to the

279 toxicity of the venom” (Ritonja et al., 1987).

280 3.1.5 Dipeptidyl peptidases

281 We find identical transcripts encoding dipeptidyl peptidase 3 and 4 in all tissues in all species

282 except the leopard gecko (Figures 2, 4a and 4b), and both of these have a low transcript

283 abundance in the venom gland of E. coloratus. dpp4 is expressed in garter snake liver and

284 Anole testis and ovary and dpp3 is also expressed in garter snake liver, king cobra pooled

285 tissues and Bearded dragon brain (Figures 4a and b). It is therefore unlikely the either dpp3 or

286 dpp4 represent venom toxins.

287 3.1.6 Epididymal secretory protein

288 We find one transcript encoding epididymal secretory protein (ESP) expressed in the venom

289 gland of Echis coloratus (9.09 FPKM) corresponding to type E1. This transcript is also found

290 to be expressed at similar levels in the scent gland (13.71 FPKM) and skin (8.64 FPKM) of

291 this species and orthologous transcripts are expressed in all three tissues of all other species

292 used in this study (Figure 2 and Supplementary figure S4a), suggesting that this is a

293 ubiquitously expressed gene and not a venom component. Previously described epididymal

294 secretory protein sequences from varanids (Fry et al., 2010) and the colubrid Cylindrophis

295 ruffus (Fry et al., 2013) do not represent esp-e1 and their true orthology is currently unclear.

296 However, our analysis of these and related sequences suggests that they are likely part of a

297 reptile-specific expansion of esp-like genes and that the Varanus and Cylindrophis sequences

298 do not encode the same gene (Supplementary figure S4b). Therefore there is not, nor was

299 there ever, any evidence that epididymal secretory protein sequences represent venom

300 components in the Toxicofera.

301 3.1.7 Ficolin (“veficolin”)

302 We find one transcript encoding ficolin in the E. coloratus venom gland and identical

303 transcripts in both scent gland and skin (Figure 2, Supplementary figure S5) and orthologous

304 transcripts in all corn snake and leopard gecko tissues, as well as rough green snake salivary

305 and scent glands and royal python salivary gland. Paralogous genes expressed in multiple

306 tissues were also found in corn snake and rough green snake (Supplementary figure S5).

307 These findings, together with additional data from available transcriptomes of pooled garter

308 snake body tissues and bearded dragon and chicken brains show that ficolin does not

309 represent a Toxicoferan venom component.

310 3.1.8 Hyaluronidase 12

311 Hyaluronidase has been suggested to be a “venom spreading factor” to aid the dispersion of

312 venom toxins throughout the body of envenomed prey, and as such it does not represent a

313 venom toxin itself (Kemparaju and Girish, 2006). We do however find two hyaluronidase

314 genes expressed in the venom gland of E. coloratus. The first appears to be venom gland

315 specific (based on available data) and has two splice variants including a truncated variant

316 similar to sequences previously characterised from Echis carinatus sochureki (accession

317 number DQ840262) and Echis pyramidum leakeyi (accession number DQ840255) venom

318 glands (Harrison et al., 2007). Although we cannot rule out hyaluronidase playing an active

319 (but non-toxic) role in Echis venom, it is worth commenting that hyaluronan has been

320 suggested to have a role in wound healing and the protection of the oral mucosa in human

321 saliva (Pogrel et al., 2003). The expression of hyaluronidases involved in hyaluronan

322 metabolism in venom and/or salivary glands is therefore perhaps unsurprising.

323 3.1.9 Kallikrein

324 We find two Kallikrein-like sequences in E. coloratus, one of which is expressed in all three

325 tissues in this species (at a low level in the venom gland) and a variety of other tissues in the

326 other study species, and one of which is found only in scent gland and skin (Figure 2,

327 Supplementary figure S6). These genes do not represent venom toxins in E. coloratus and

328 appear to be most closely-related to a group of mammalian Kallikrein (KLK) genes

329 containing KLK1, 11, 14 and 15 and probably represent the outgroup to a mammalian-

330 specific expansion of this gene family. The orthology of previously published Toxicoferan

331 Kallikrein genes is currently unclear and the majority of these sequences can be found in our

332 serine protease tree (see later section and Supplementary figure S19).

333 3.1.10 Kunitz

334 We find a number of transcripts encoding Kunitz-type protease inhibitors in our tissue data,

335 with the majority of these encoding kunitz1 and kunitz2 genes (Figure 2 and Supplementary

336 figure S7). The tissue distribution of these transcripts, together with the phylogenetic position 13

337 of lizard and venomous snake sequences does not support a monophyletic clade of venom

338 gland-specific Kunitz-type genes in the Toxicofera. The presence of protease inhibitors in

339 reptile venom and salivary glands should perhaps not be too surprising and it again seems

340 likely that the involvement of Kunitz-type inhibitors in venom toxicity in some advanced

341 snake lineages (in this case mamba (Dendroaspis spp.) dendrotoxins and krait (Bungarus

342 multicinctus) bungarotoxins (Harvey, 2001; Kwong et al., 1995)) has led to confusion when

343 non-toxic orthologs have been identified in other species.

344 3.1.11 Lysosomal acid lipase

345 We find two transcripts encoding Lysosomal acid lipase genes in the E. coloratus venom

346 gland transcriptome, one of which (lipa-a) is also expressed in skin and scent gland in this

347 species and all three tissues in our other study species. lipa-a, despite not being venom gland

348 specific, is more highly expressed in the venom gland (3,337.33 FPKM) than in the scent

349 gland (484.49 FPKM) and skin (22.79 FPKM) of E. coloratus, although there is no evidence

350 of elevated expression in the salivary glands of our other study species. As this protein is

351 involved in lysosomal lipid hydrolysis (Warner et al., 1981) and the venom gland is a highly

352 active tissue, we suggest that this elevated expression is likely related to high cell turnover.

353 Transcripts of lipa-b are found at a low level in the venom and scent glands of E. coloratus

354 and the scent gland of royal python (Figure 2, Supplementary figure S8). Neither lipa-a or

355 lipa-b therefore encode venom toxins.

356 3.1.12 Natriuretic peptide

357 We find only a single natriuretic peptide-like sequence in our dataset, in the skin of the royal

358 python. The absence of this gene from the rest of our study species suggests that it is not a

359 highly conserved Toxicoferan toxin.

360 3.1.13 Nerve growth factor

361 We find identical transcripts encoding nerve growth factor (ngf) in all three E. coloratus

362 tissues. Transcripts encoding the orthologous gene are also found in the corn snake salivary 14

363 gland and scent gland; rough green snake scent gland and skin; royal python skin and leopard

364 gecko salivary gland, scent gland and skin (Figure 2 and Supplementary figure S9). ngf is

365 expressed at a higher level in the venom gland (525.82 FPKM) than in the scent gland (0.18

366 FPKM) and skin (0.58 FPKM) of E. coloratus, but not at an elevated level in the salivary

367 gland of other species, again hinting at the potential for exaptation of this gene. Based on

368 these findings, together with the expression of this gene in garter snake pooled tissues, we

369 suggest that ngf does not encode a Toxicoferan toxin. However, we do find evidence for the

370 duplication of ngf in cobras (Supplementary figure S9), suggesting that it may represent a

371 venom toxin in at least some advanced snakes (Sunagar et al., 2013). As with complement c3,

372 it seems likely that the identification of non-toxic orthologs in distantly-related species has

373 led to the conclusion that ngf is a widely-distributed venom toxin and confused its true

374 evolutionary history.

375 3.1.14 Phospholipase A2 (PLA2 Group IIE)

376 We find transcripts encoding Group IIE PLA2 genes in the venom gland of E. coloratus and

377 the salivary glands of all other species (Figure 2, Supplementary figure S10). Although this

378 gene appears to be venom and salivary-gland-specific (based on available data), its presence

379 in all species (including the non-Toxicoferan leopard gecko) suggests that it does not

380 represent a toxic venom component.

381 3.1.15 Phospholipase B

382 We find a single transcript encoding phospholipase b expressed in all three E. coloratus

383 tissues (Figures 2 and 5) and transcripts encoding the orthologous gene are found in all other

384 tissues from all study species, with the exception of rough green snake salivary gland. We

385 also find plb transcripts in corn snake vomeronasal organ, garter snake liver, Burmese python

386 pooled tissues (liver and heart) and bearded dragon brain (Figure 5). The two transcripts in

387 the rough green snake and corn snake are likely alleles or the result of individual variation,

388 and actually represent a single phospholipase b gene from each of these species. Transcript 15

389 abundance analysis shows this gene to be expressed at a low level in all tissues from all study

390 species. Based on the phylogenetic and tissue distribution of this gene it is unlikely to

391 represent a Toxicoferan venom toxin.

392 3.1.16 Renin (“renin aspartate protease”)

393 We find a number of transcripts encoding renin-like genes in the E. coloratus venom gland

394 (Figures 2 and 6), one of which (encoding the canonical renin) is also expressed in the scent

395 gland and is orthologous to a previously described sequence from the venom gland of the

396 ocellated carpet viper (Echis ocellatus, accession number CAJ55260). We also find that the

397 recently-published Boa constrictor renin aspartate protease (rap) gene (accession number

398 JX467165 (Fry et al., 2013)) is in fact a cathepsin d gene, transcripts of which are found in all

399 three tissues in all five of our study species. We suggest that this misidentification may be

400 due to a reliance on BLAST-based classification, most likely using a database restricted to

401 squamate or serpent sequences. It is highly unlikely that either renin or cathepsin d (or indeed

402 any renin-like aspartate proteases) constitute venom toxins in E. coloratus or E. ocellatus, nor

403 do they represent basal Toxicoferan toxins.

404 3.1.17 Ribonuclease

405 Ribonucleases have been suggested to have a role in the generation of free purines in snake

406 venoms (Aird, 2005) and the presence of these genes in the salivary glands of two species of

407 lizard (Gerrhonotus infernalis and Celestus warreni) and two colubrid snakes (Liophis

408 peocilogyrus and Psammophis mossambicus) has been used to support the Toxicofera (Fry et

409 al., 2010; Fry et al., 2012b). We did not identify orthologous ribonuclease genes in any of our

410 salivary or venom gland data, nor do we find them in venom gland transcriptomes from the

411 Eastern diamondback rattlesnake, king cobra and eastern coral snake (although we have

412 identified a wide variety of other ribonuclease genes). The absence of these genes in seven

413 Toxicoferans, coupled with the fact that they were initially described from only 2 out of 11

414 species of snake (Fry et al., 2012b) and 3 out of 18 species of lizard (Fry et al., 2010) would

415 cast doubt on their status as conserved Toxicoferan toxins.

416 3.1.18 Three finger toxins (3ftx)

417 We find 2 transcripts encoding three finger toxin (3ftx)-like genes expressed in the E.

418 coloratus venom gland, one of which is expressed in all 3 tissues (3ftx-a) whilst the other is

419 expressed in the venom and scent glands (3ftx-b). Orthologous transcripts of 3ftx-a are found

420 to be expressed in all three tissues of corn snake, rough green snake salivary gland and skin,

421 and royal python salivary gland. An ortholog of 3ftx-b is expressed in rough green snake

422 scent gland. We also find a number of different putative 3ftx genes in our other study species,

423 often expressed in multiple tissues (Figure 2, Supplementary figure S11). Based on the

424 phylogenetic and tissue distribution of both of these genes we suggest that they do not

425 represent venom toxins in E. coloratus. As with other proposed Toxicoferan genes such as

426 complement c3 and nerve growth factor, it seems likely that 3ftx genes are indeed venom

427 components in some species, especially cobras and other elapids (Vonk et al., 2013; Fry et

428 al., 2003), and that the identification of their non-venom orthologs in other species has led to

429 much confusion regarding the phylogenetic distribution of these toxic variants.

430 3.1.19 Vespryn

431 We do not find vespryn transcripts in any E. coloratus tissues, although this gene is present in

432 the genome of this species (accession number KF114032). We do however find transcripts

433 encoding this gene in the salivary and scent glands of the corn snake, and skin and scent glands

434 of the rough green snake, royal python and leopard gecko (Figure 2, Supplementary figure

435 S12). We suggest that the tissue distribution of this gene in these species casts doubt on its role

436 as a venom component in the Toxicofera.

437 3.1.20 Waprin

438 We find a number of “waprin”-like genes in our dataset, expressed in a diverse array of body

439 tissues. Our phylogenetic analyses (Supplementary figure S13) show that previously 17

440 characterised “waprin” genes (Torres et al., 2003; Fry et al., 2008; Rokyta et al., 2012; Aird

441 et al., 2013; Nair et al., 2007) most likely represent WAP four-disulfide core domain 2

442 (wfdc2) genes, which have undergone a squamate-specific expansion for which there is no

443 evidence for a venom gland-specific paralog. It is unlikely therefore that these genes

444 represent a Toxicoferan venom toxin. Indeed, the inland taipan (Oxyuranus microlepidotus)

445 “Omwaprin” has been shown to be “…non-toxic to Swiss albino mice at doses of up to 10

446 mg/kg when administered intraperitoneally” (Nair et al., 2007) and is more likely to have an

447 antimicrobial function in the venom or salivary gland.

448 3.2 Putative venom toxins of Echis coloratus

449 The following genes show either a venom gland-specific expression or an elevated expression

450 level in this tissue, but not both. As such we suggest that whilst they may represent venom

451 toxins in E. coloratus, further analysis is needed in order to confirm this.

452 3.2.1 Vascular endothelial growth factor

453 We find four transcripts encoding vascular endothelial growth factor (VEGF) expressed in the

454 venom gland of E. coloratus. These correspond to vegf-a, vegf-b, vegf-c and vegf-f and of these,

455 vegf-a, b and c are also expressed in the skin and scent gland of this species (Figure 2).

456 Transcripts encoding orthologs of these genes are expressed in all three tissues of all other

457 species used in this study (with the exception of the absence of vegf-a in corn snake skin). In

458 accordance with previous studies (Rokyta et al., 2011), we find evidence of alternative splicing

459 of vegf-a transcripts in all species although no variant appears to be tissue-specific. It is likely

460 that a failure to properly recognise and classify alternatively spliced vegf-a transcripts (Aird et

461 al., 2013) may have contributed to an overestimation of snake venom complexity. vegf-d was

462 only found to be expressed in royal python salivary gland and scent gland and all three tissues

463 from leopard gecko (Figure 2, Supplementary figure S14). The transcript encoding VEGF-F is

464 found only in the venom gland of E. coloratus and, given the absence of any Elapid vegf-f

465 sequences in public databases as well as absence of this transcript in the two species of colubrid 18

466 in our study, we suggest that vegf-f is specific to vipers. Whilst vegf-f has a higher transcript

467 abundance in E. coloratus venom gland (186.73 FPKM) than vegf-a (3.24 FPKM), vegf-b (1.28

468 FPKM) and vegf-c (1.54 FPKM), compared to other venom genes in this species (see next

469 section) it has a considerably lower transcript abundance suggesting it represents at most a

470 minor venom component in E. coloratus.

471 3.2.2 L-amino acid oxidase

472 We find transcripts encoding two l-amino acid oxidase (laao) genes in E. coloratus, one of

473 which (laao-b) has two splice variants (Figure 2, Supplementary figure S15). laao-a

474 transcripts are found in all three E. coloratus and leopard gecko tissues. laao-b is venom

475 gland-specific in E. coloratus (based on the available data) and transcripts of the orthologous

476 gene are found in the scent glands of corn snake, rough green snake and royal python. The

477 splice variant laao-b2 may represent a venom toxin in E. coloratus based on its specific

478 expression in the venom gland of this species and elevated expression level (628.84 FPKM).

479 3.2.3 Crotamine

480 We find a single crotamine-like transcript in the venom gland of E. coloratus (Figure 2).

481 Related genes are found in a variety of tissues in other study species (including the scent

482 gland of the rough green snake, the salivary gland and skin of the leopard gecko, and in all

483 three corn snake tissues), although the short length of these sequences precludes a definitive

484 statement of orthology. This gene may represent a toxic venom component in E. coloratus

485 based on its tissue distribution, but due to its low transcript abundance (10.95 FPKM) it is

486 likely to play a minor role, if any.

487 3.3 Proposed venom toxins in Echis coloratus

488 The following genes are found only in the venom gland of E. coloratus and clearly show an

489 elevated expression level (Figure 7). Whilst we classify these genes as encoding venom

490 toxins in this species (Table 1) it should be noted that none of these genes support the

491 monophyly of Toxicoferan venom toxins. 19

492 3.3.1 Cysteine-rich secretory proteins (CRISPs)

493 We find transcripts encoding two distinct CRISPs expressed in the E. coloratus venom gland,

494 one of which is also found in skin and scent gland (Figure 2). Phylogenetic analysis of these

495 genes (which we call crisp-a and crisp-b) reveals that they appear to have been created as a

496 result of a gene duplication event earlier in the evolution of advanced snakes (Supplementary

497 Figure S16). crisp-a transcripts are also found in all three corn snake tissues, as well as rough

498 green snake skin and scent gland and royal python scent gland. crisp-b is also found in corn

499 snake salivary gland (Figure 2 and Supplementary figure S16) and the phylogenetic and

500 tissue distribution of this gene suggest that it does indeed represent a venom toxin, produced

501 via duplication of an ancestral crisp gene that was expressed in multiple tissues, including the

502 salivary gland. The elevated transcript abundance of crisp-b (3,520.07 FPKM) in the venom

503 gland of E. coloratus further supports its role as a venom toxin in this species (Figure 7). The

504 phylogenetic and tissue distribution and low transcript abundance of crisp-a (0.61 FPKM in

505 E. coloratus venom gland) shows that it is unlikely to be a venom toxin. We also find no

506 evidence of a monophyletic clade of reptile venom toxins and therefore suggest that, contrary

507 to earlier reports (Fry et al., 2009b; Fry et al., 2010), the CRISP genes of varanid and

508 helodermatid lizards do not represent shared Toxicoferan venom toxins and, if they are

509 indeed toxic venom components, they have been recruited independently from those of the

510 advanced snakes. Regardless of their status as venom toxins, it appears likely that the

511 diversity of CRISP genes in varanid lizards in particular (Fry et al., 2006) has been

512 overestimated as a result of the use of negligible levels of sequence variation to classify

513 transcripts as representing distinct gene products (Supplementary figures S23 and S24).

514 3.3.2 C-type lectins

515 We find transcripts encoding 11 distinct C-type lectin genes in the E. coloratus venom gland,

516 one of which (ctl-a) is also expressed in the scent gland of this species. The remaining 10

517 genes (ctl-b to k) are found only in the venom gland and form a clade with other viper C-type 20

518 lectin genes (Figure 2, Supplementary figure S17). Of these, 6 are highly expressed in the

519 venom gland (ctl-b to d, ctl-f to g and ctl-j) with a transcript abundance range of 3,706.21-

520 24,122.41 FPKM (Figure 7). The remainder of these genes (ctl-e, ctl-h to i and ctl-k) show

521 lower transcript abundance (0.80-1,475.88 FPKM), with two (ctl-i and k) being more lowly

522 expressed than ctl-a (230.06 FPKM). A number of different C-type lectin genes are found in

523 our other study species, often expressed in multiple tissues (Supplementary figure S17). We

524 therefore suggest that the 6 venom-gland specific C-type lectin genes that are highly

525 expressed are indeed venom toxins in E. coloratus and that these genes diversified via the

526 duplication of an ancestral gene with a wide expression pattern, including in salivary/venom

527 glands. Based on their selective expression in the venom gland (from available data) the

528 remaining four C-type lectin genes cannot be ruled out as putative toxins, although their

529 lower transcript abundance suggests that they are likely to be minor components in E.

530 coloratus venom. It should also be noted that a recent analysis of king cobra (Ophiohagus

531 hannah) venom gland transcriptome and proteome suggested that “…lectins do not contribute

532 to king cobra envenoming” (Vonk et al., 2013).

533 3.3.3 Phospholipase A2 (PLA2 Group IIA)

534 We find five transcripts encoding Group IIA PLA2 genes in E. coloratus, three of which are

535 found only in the venom gland and two of which are found only in the scent gland (these

536 latter two likely represent intra-individual variation in the same transcript) (Figure 2,

537 Supplementary figure S18). The venom gland-specific transcript PLA2 IIA-c is highly

538 expressed (22,520.41 FPKM) and likely represents a venom toxin, and may also be a putative

539 splice variant although further analysis is needed to confirm this. PLA2 IIA-d and IIA-e show

540 an elevated, but lower, expression level (1,677.15 FPKM and 434.67 FPKM respectively,

541 Figure 7). Based on tissue and phylogenetic distribution we would propose that these three

542 genes may represent putative venom toxins (Table 1).

543 3.3.4 Serine proteases 21

544 We find 6 transcripts encoding Serine proteases in E. coloratus (Figure 2, Supplementary

545 figure S19) which (based on available data) are all venom gland specific. Four of these

546 transcripts are highly expressed in the venom gland (serine proteases a-c and e; 3,076.01-

547 7,687.03 FPKM) whilst two are expressed at a lower level (serine proteases d and f; 1,098.45

548 FPKM and 102.34 FPKM respectively, Figure 7). Based on these results we suggest serine

549 proteases a, b, c and e represent venom toxins whilst serine proteases d and f may represent

550 putative venom toxins (Table 1).

551 3.3.5 Snake venom metalloproteinases

552 We find 21 transcripts encoding snake venom metalloproteinases in E. coloratus and of these

553 14 are venom gland-specific, whilst another (svmp-n) is expressed in the venom gland and

554 scent gland. Five remaining genes are expressed in the scent gland only whilst another is

555 expressed in the skin (Figure 2, Supplementary figure S20). Of the 14 venom gland-specific

556 SVMPs we find 4 to be highly expressed (5,552.84-15,118.41 FPKM, Figure 7). In the

557 absence of additional data, we classify the 13 venom gland-specific svmp genes as venom

558 toxins in this species (Table 1).

559 4. Discussion

560 Our transcriptomic analyses have revealed that all 16 of the basal venom toxin genes used to

561 support the hypothesis of a single, early evolution of venom in reptiles (the Toxicofera

562 hypothesis (Vidal and Hedges, 2005; Fry et al., 2006; Fry et al., 2009a; Fry et al., 2009b; Fry

563 et al., 2010; Fry et al., 2012a; Fry et al., 2013)), as well as a number of other genes that have

564 been proposed to encode venom toxins in multiple species are in fact expressed in multiple

565 tissues, with no evidence for consistently higher expression in venom or salivary glands.

566 Additionally, only two genes in our entire dataset of 74 genes in five species were found to

567 encode possible venom gland-specific splice variants (l-amino acid oxidase b2 and PLA2 IIA-

568 c). We therefore suggest that many of the proposed basal Toxicoferan genes most likely

569 represent housekeeping or maintenance genes and that the identification of these genes as 22

570 conserved venom toxins is a side-effect of incomplete tissue sampling. This lack of support

571 for the Toxicofera hypothesis therefore prompts a return to the previously held view

572 (Kardong et al., 2009) that venom in different lineages of reptiles has evolved independently,

573 once at the base of the advanced snakes, once in the helodermatid (gila monster and beaded

574 lizard) lineage and, possibly, one other time in monitor lizards, although evidence for a

575 venom system in this latter group (Fry et al., 2009b; Fry et al., 2010; Vikrant and Verma,

576 2013) may need to be reinvestigated in light of our findings. The process of reverse

577 recruitment (Casewell et al., 2012), where a venom gene undergoes additional gene

578 duplication events and is subsequently recruited from the venom gland back into a body

579 tissue (which was proposed on the basis of the placement of garter snake and Burmese

580 python “physiological” genes within clades of “venom” genes) must also be re-evaluated in

581 light of our findings.

582 Bites by venomous snakes are thought to be responsible for as many as 1,841,000

583 envenomings and 94,000 deaths annually (predominantly in the developing world

584 (Kasturiratne et al., 2008; Harrison et al., 2009)), and medical treatment of snakebite is reliant

585 on the production of antivenoms containing antibodies, typically from sheep or horses, that

586 will bind and neutralise toxic venom proteins (Chippaux and Goyffon, 1998). Since these

587 antivenoms are derived from the injection of crude venom into the host animal, they are not

588 targeted to the most pathogenic venom components and therefore also include antibodies to

589 weakly- or non-pathogenic proteins requiring the administration of large or multiple doses

590 (Casewell et al., 2013), increasing the risks of adverse reactions. A comprehensive

591 understanding of snake venom composition is therefore vital for the development of the next

592 generation of antivenoms (Harrison, 2004; Wagstaff et al., 2006; Casewell et al., 2013) as it

593 is important that research effort is not spread too thinly through the inclusion of non-toxic

594 venom gland transcripts. Our results suggest that erroneous assumptions about the single

595 origination and functional conservation of venom toxins across the Toxicofera has led to the 23

596 complexity of snake venom being overestimated by previous authors. We propose that the

597 venom of the painted saw-scaled viper, Echis coloratus, is likely to consist of just 34 genes in

598 8 gene families (Table 1, based on venom gland-specific expression and a ‘high’ expression,

599 as defined by presence in the top 25% of transcripts (Williford and Demuth, 2012) in at least

600 two of four venom gland samples), fewer than has been suggested for this and related species

601 in previous EST or transcriptomic studies (Wagstaff and Harrison, 2006; Casewell et al.,

602 2009).

603 It is noteworthy that the results of our analyses accord well with proteomic analyses of

604 venom composition in snakes, with an almost identical complement of 35 toxins in 8 gene

605 families known from the related ocellated carpet viper, Echis ocellatus (Wagstaff et al.,

606 2009), where SVMPs, CTLs and PLA2s were found to be the most abundant proteins. Studies

607 of a range of other venomous snake species have identified a typical complement of between

608 24-61 toxins in 6-14 families (Table 2). Far from being a “complex cocktail” (Izidoro et al.,

609 2006; Calvete et al., 2007b; Wong and Belov, 2012; Casewell et al., 2013), snake venom may

610 in fact represent a relatively simple mixture of toxic proteins honed by natural selection for

611 rapid prey immobilisation, with limited lineage-specific expansion in one or a few particular

612 gene families.

613 In order to avoid continued overestimation of venom complexity, we propose that future

614 transcriptome-based analyses of venom composition must include quantitative comparisons

615 of multiple body tissues from multiple individuals and robust phylogenetic analysis that

616 includes known paralogous members of gene families. We would also encourage the use of

617 clearly explained, justifiable criteria for classifying highly similar sequences as new paralogs

618 rather than alleles or the result of PCR or sequencing errors, as it seems likely that some

619 available sequences from previous studies have been presented as distinct genes on the basis

620 of extremely minor (or even non-existent) sequence variation (see Supplementary figures

621 S21-S24 for examples of identical or nearly identical ribonuclease and CRISP sequences and 24

622 Supplementary figures S25 and S26 for examples of the same sequence being annotated as

623 two different genes). As a result, the diversity of “venom” composition in these species may

624 have been inadvertently inflated.

625 Additionally, we would encourage the adoption of a standard nomenclature for reptile genes,

626 as the overly-complicated and confusing nomenclature used currently (Table 3) may also

627 contribute to the perceived complexity of snake venom. We propose that such a nomenclature

628 system should be based on the comprehensive standards developed for anole lizards (Kusumi

629 et al., 2011; Hargreaves & Mulley, 2014b). It seems likely that the application of our

630 approach to other species (together with proteomic studies of extracted venom) will lead to a

631 commensurate reduction in claimed venom diversity, with clear implications for the

632 development of next generation antivenoms: since most true venom genes are members of a

633 relatively small number of gene families, it is likely that a similarly small number of

634 antibodies may be able to bind to and neutralise the toxic venom components, especially with

635 the application of “string of beads” techniques (Whitton et al., 1993) utilising fusions of short

636 oligopeptide epitopes designed to maximise the cross-reactivity of the resulting antibodies

637 (Wagstaff et al., 2006).

638

639 5. Conclusions

640 We suggest that identification of the apparently conserved Toxicofera venom toxins in

641 previous studies is most likely a side effect of incomplete tissue sampling, compounded by

642 incorrect interpretation of phylogenetic trees and the use of BLAST-based gene identification

643 methods. It should perhaps not be too surprising that homologous tissues in related species

644 would show similar gene complements and the restriction of most previous studies to only the

645 “venom” glands means that monophyletic clades of reptile sequences in phylogenetic trees

646 have been taken to represent monophyletic clades of venom toxin genes. Whilst it is true that

647 some of these genes do encode toxic proteins in some species (indeed, this was often the 25

648 basis for their initial discovery) the discovery of orthologous genes in other species does not

649 necessarily demonstrate shared toxicity. In short, toxicity in one does not equal toxicity in all.

650

651 Acknowledgements

652 The authors wish to thank A. Tweedale, R. Morgan, G. Cooke, A. Barlow, C. Wüster and M.

653 Hegarty for technical assistance and N. Casewell, W. Wüster and A. Malhotra for useful

654 discussions. We are grateful to the staff of High Performance Computing (HPC) Wales for

655 enabling and supporting our access to their systems, and to Richard Storey and Daniel

656 Struthers of PetGen Ltd. for their partnership during this project. This research was supported

657 by a Royal Society Research Grant awarded to JFM (grant number RG100514) and

658 Wellcome trust funding to DWL (grant number 098051). JFM and MTS are supported by the

659 Biosciences, Environment and Agriculture Alliance (BEAA) between Bangor University and

660 Aberystwyth University and ADH is funded by a Bangor University 125th Anniversary

661 Studentship.

662

663

664 References

665 Aird, S.D., 2008. Snake venom dipeptidyl peptidase IV: taxonomic distribution and

666 quantitative variation. Comp. Biochem. Physiol. B. Biochem. Mol. Biol. 150, 222-228.

667

668 Aird, S.D., 2005. Taxonomic distribution and quantitative analysis of free purine and

669 pyrimidine nucleosides in snake venoms. Comp. Biochem. Physiol. B. Biochem. Mol. Biol.

670 140, 109-126.

671

672 Aird, S.D., Watanabe, Y., Villar-Briones, A., Roy, M.C., Terada, K., Mikheyev, A.S., 2013.

673 Quantitative high-throughput profiling of snake venom gland transcriptomes and proteomes

674 (Ovophis okinavensis and Protobothrops flavoviridis). BMC Genomics 14, 790.

675

676 Alper, C., Balavitch, D., 1976. Cobra venom factor: evidence for its being altered cobra C3

677 (the third component of complement). Science 191, 1275-1276.

678

679 Andrews, S., 2010. FastQC: A quality control tool for high throughput sequence data.

680 Available online at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc.

681

682 Angeletti, R.H., 1970. Nerve growth factor from cobra venom. Proc. Natl. Acad. Sci. U. S. A.

683 65, 668-674.

684

685 Bernheimer, A., Linder, R., Weinstein, S., Kim, K., 1987. Isolation and characterization of a

686 phospholipase B from venom of collett's snake, Pseudechis colletti. Toxicon 25, 547-554.

687

688 Bjarnason, J.B., Fox, J.W., 1994. Hemorrhagic metalloproteinases from snake venoms.

689 Pharmacol. Ther. 62, 325-372. 27

690

691 Brykczynska, U., Tzika, A.C., Rodriguez, I., Milinkovitch, M.C., 2013. Contrasted evolution

692 of the vomeronasal receptor repertoires in mammals and squamate reptiles. Genome Biol.

693 Evol. 5, 389-401.

694

695 Butte, A.J., Dzau, V.J., Glueck, S.B., 2002. Further defining housekeeping, or" maintenance,"

696 genes: Focus on" A compendium of gene expression in normal human tissues". Physiological

697 genomics 7, 95-95.

698

699 Calvete, J.J., Escolano, J., Sanz, L., 2007a. Snake venomics of Bitis species reveals large

700 intragenus venom toxin composition variation: application to taxonomy of congeneric taxa. J.

701 Proteome Res. 6, 2732-2745.

702

703 Calvete, J.J., Fasoli, E., Sanz, L., Boschetti, E., Righetti, P.G., 2009. Exploring the venom

704 proteome of the Western diamondback rattlesnake, Crotalus atrox, via snake venomics and

705 combinatorial peptide ligand library approaches. J. Proteome Res. 8, 3055-3067.

706

707 Calvete, J.J., Juárez, P., Sanz, L., 2007b. Snake venomics. Strategy and applications. J. Mass

708 Spectrom. 42, 1405-1414.

709

710 Calvete, J.J., Marcinkiewicz, C., Sanz, L., 2006. Snake venomics of Bitis gabonica gabonica.

711 Protein family composition, subunit organization of venom toxins, and characterization of

712 dimeric disintegrins bitisgabonin-1 and bitisgabonin-2. J. Proteome Res. 6, 326-336.

713

714 Casewell, N.R., Huttley, G.A., Wüster, W., 2012. Dynamic evolution of venom proteins in

715 squamate reptiles. Nat. Commun. 3, 1066. 28

716

717 Casewell, N.R., Wüster, W., Vonk, F.J., Harrison, R.A., Fry, B.G., 2013. Complex cocktails:

718 the evolutionary novelty of venoms. Trends Ecol. Evol. 28, 219-229.

719

720 Casewell, N., Harrison, R., Wüster, W., Wagstaff, S., 2009. Comparative venom gland

721 transcriptome surveys of the saw-scaled vipers (Viperidae: Echis) reveal substantial intra-

722 family gene diversity and novel venom transcripts. BMC Genomics 10, 564.

723

724 Castoe, T.A., de Koning, A.P., Hall, K.T., Card, D.C., Schield, D.R., Fujita, M.K., Ruggiero,

725 R.P., Degner, J.F., Daza, J.M., Gu, W., Reyes-Velasco, J., Shaney, K.J., Castoe, J.M., Fox,

726 S.E., Poole, A.W., Polanco, D., Dobry, J., Vandewege, M.W., Li, Q., Schott, R.K., Kapusta,

727 A., Minx, P., Feschotte, C., Uetz, P., Ray, D.A., Hoffmann, F.G., Bogden, R., Smith, E.N.,

728 Chang, B.S., Vonk, F.J., Casewell, N.R., Henkel, C.V., Richardson, M.K., Mackessy, S.P.,

729 Bronikowski, A.M., Yandell, M., Warren, W.C., Secor, S.M., Pollock, D.D., 2013. The

730 Burmese python genome reveals the molecular basis for extreme adaptation in snakes. Proc.

731 Natl. Acad. Sci. U. S. A. 110, 20645-20650.

732

733 Castoe, T.A., Fox, S.E., Jason de Koning, A., Poole, A.W., Daza, J.M., Smith, E.N., Mockler,

734 T.C., Secor, S.M., Pollock, D.D., 2011. A multi-organ transcriptome resource for the

735 Burmese Python (Python molurus bivittatus). BMC Res. Notes 4, 310-0500-4-310.

736

737 Chatrath, S.T., Chapeaurouge, A., Lin, Q., Lim, T.K., Dunstan, N., Mirtschin, P., Kumar,

738 P.P., Kini, R.M., 2011. Identification of novel proteins from the venom of a cryptic snake

739 Drysdalia coronoides by a combined transcriptomics and proteomics approach. J. Proteome

740 Res. 10, 739-750.

741 29

742 Chippaux, J., Goyffon, M., 1998. Venoms, antivenoms and immunotherapy. Toxicon 36,

743 823-846.

744

745 Cotton, J.A., Page, R.D., 2005. Rates and patterns of gene duplication and loss in the human

746 genome. Proc. Biol. Sci. 272, 277-283.

747

748 Cousin, X., Créminon, C., Grassi, J., Méflah, K., Cornu, G., Saliou, B., Bon, S., Massoulié,

749 J., Bon, C., 1996a. Acetylcholinesterase from Bungarus venom: a monomeric species. FEBS

750 Lett. 387, 196-200.

751

752 Cousin, X., Bon, S., Duval, N., Massoulie, J., Bon, C., 1996b. Cloning and expression of

753 acetylcholinesterase from Bungarus fasciatus venom. A new type of COOH-terminal

754 domain; involvement of a positively charged residue in the peripheral site. J. Biol. Chem.

755 271, 15099-15108.

756

757 Cousin, X., Bon, S., Massoulie, J., Bon, C., 1998. Identification of a novel type of

758 alternatively spliced exon from the acetylcholinesterase gene of Bungarus fasciatus.

759 Molecular forms of acetylcholinesterase in the snake liver and muscle. J. Biol. Chem. 273,

760 9812-9820.

761

762 Junqueira de Azevedo, I.L.M., Farsky, S.H.P., Oliveira, M.L.S., Ho, P.L., 2001. Molecular

763 Cloning and Expression of a Functional Snake Venom Vascular Endothelium Growth Factor

764 (VEGF) from the Bothrops insularis Pit Viper: a new member of the VEGF family of

765 proteins. J. Biol. Chem. 276, 39836-39842.

766

767 Du, X., Clemetson, K.J., 2002. Snake venom L-amino acid oxidases. Toxicon 40, 659-665. 30

768

769 Eckalbar, W.L., Hutchins, E.D., Markov, G.J., Allen, A.N., Corneveaux, J.J., Lindblad-Toh,

770 K., Di Palma, F., Alfoldi, J., Huentelman, M.J., Kusumi, K., 2013. Genome reannotation of

771 the lizard Anolis carolinensis based on 14 adult and embryonic deep transcriptomes. BMC

772 Genomics 14, 49-2164-14-49.

773

774 Fahmi, L., Makran, B., Pla, D., Sanz, L., Oukkache, N., Lkhider, M., Harrison, R.A., Ghalim,

775 N., Calvete, J.J., 2012. Venomics and antivenomics profiles of North African Cerastes

776 cerastes and C. vipera populations reveals a potentially important therapeutic weakness.

777 Journal of proteomics 75, 2442-2453.

778

779 Fry, B.G., Wüster, W., Kini, R.M., Brusic, V., Khan, A., Venkataraman, D., Rooney, A.,

780 2003. Molecular evolution and phylogeny of elapid snake venom three-finger toxins. J. Mol.

781 Evol. 57, 110-129.

782

783 Fry, B.G., Casewell, N.R., Wüster, W., Vidal, N., Young, B., Jackson, T.N., 2012a. The

784 structural and functional diversification of the Toxicofera reptile venom system. Toxicon 60,

785 434-448.

786

787 Fry, B.G., Scheib, H., Junqueira de Azevedo, Inacio de LM, Silva, D.A., Casewell, N.R.,

788 2012b. Novel transcripts in the maxillary venom glands of advanced snakes. Toxicon 59,

789 696-708.

790

791 Fry, B.G., Undheim, E.A., Ali, S.A., Debono, J., Scheib, H., Ruder, T., Jackson, T.N.,

792 Morgenstern, D., Cadwallader, L., Whitehead, D., 2013. Squeezers and leaf-cutters:

793 differential diversification and degeneration of the venom system in toxicoferan reptiles. Mol.

794 Cell. Proteomics 12, 1881-1899.

795

796 Fry, B.G., Vidal, N., Norman, J.A., Vonk, F.J., Scheib, H., Ramjan, S.R., Kuruppu, S., Fung,

797 K., Hedges, S.B., Richardson, M.K., 2006. Early evolution of the venom system in lizards

798 and snakes. Nature 439, 584-588.

799

800 Fry, B.G., Vidal, N., Van der Weerd, L., Kochva, E., Renjifo, C., 2009a. Evolution and

801 diversification of the Toxicofera reptile venom system. Journal of proteomics 72, 127-136.

802

803 Fry, B.G., Wroe, S., Teeuwisse, W., van Osch, M.J., Moreno, K., Ingle, J., McHenry, C.,

804 Ferrara, T., Clausen, P., Scheib, H., 2009b. A central role for venom in predation by Varanus

805 komodoensis (Komodo dragon) and the extinct giant Varanus (Megalania) priscus. Proc.

806 Natl. Acad. Sci. U. S. A. 106, 8969-8974.

807

808 Fry, B.G., 2005. From genome to "venome": molecular origin and evolution of the snake

809 venom proteome inferred from phylogenetic analysis of toxin sequences and related body

810 proteins. Genome Res. 15, 403-420.

811

812 Fry, B.G., Scheib, H., van der Weerd, L., Young, B., McNaughtan, J., Ramjan, S.F., Vidal,

813 N., Poelmann, R.E., Norman, J.A., 2008. Evolution of an arsenal: structural and functional

814 diversification of the venom system in the advanced snakes (Caenophidia). Mol. Cell.

815 Proteomics 7, 215-246.

816

817 Fry, B.G., Winter, K., Norman, J.A., Roelants, K., Nabuurs, R.J., van Osch, M.J., Teeuwisse,

818 W.M., van der Weerd, L., McNaughtan, J.E., Kwok, H.F., Scheib, H., Greisman, L., Kochva, 32

819 E., Miller, L.J., Gao, F., Karas, J., Scanlon, D., Lin, F., Kuruppu, S., Shaw, C., Wong, L.,

820 Hodgson, W.C., 2010. Functional and structural diversification of the Anguimorpha lizard

821 venom system. Mol. Cell. Proteomics 9, 2369-2390.

822

823 Georgieva, D., Risch, M., Kardas, A., Buck, F., von Bergen, M., Betzel, C., 2008.

824 Comparative analysis of the venom proteomes of Vipera ammodytes ammodytes and Vipera

825 ammodytes meridionalis. J. Proteome Res. 7, 866-886.

826

827 Gould, S.J., Vrba, E.S., 1982. Exaptation-a missing term in the science of form. Paleobiology

828 8, 4-15.

829

830 Grabherr, M.G., Haas, B.J., Yassour, M., Levin, J.Z., Thompson, D.A., Amit, I., Adiconis,

831 X., Fan, L., Raychowdhury, R., Zeng, Q., 2011. Full-length transcriptome assembly from

832 RNA-Seq data without a reference genome. Nat. Biotechnol. 29, 644-652.

833

834 Hargreaves, A.D., Swain, M.T., Hegarty, M.J., Logan, D.W., Mulley, J.F. (2014a) Restriction

835 and recruitment – gene duplication and the origin and evolution of snake venom toxins

836 Genome Biology & Evolution 6, 2088-2095

837

838 Hargreaves, A.D. Mulley, J.F. (2014b) A plea for standardized nomenclature of snake venom

839 toxins. Toxicon (in press) doi: 10.1016/j.toxicon.2014.08.070.

840

841 Harrison, R., 2004. Development of venom toxin-specific antibodies by DNA immunisation:

842 rationale and strategies to improve therapy of viper envenoming. Vaccine 22, 1648-1655.

843

844 Harrison, R.A., Hargreaves, A., Wagstaff, S.C., Faragher, B., Lalloo, D.G., 2009. Snake

845 envenoming: a disease of poverty. PLoS Negl. Trop. Dis. 3, e569.

846

847 Harrison, R.A., Ibison, F., Wilbraham, D., Wagstaff, S.C., 2007. Identification of cDNAs

848 encoding viper venom hyaluronidases: cross-generic sequence conservation of full-length and

849 unusually short variant transcripts. Gene 392, 22-33.

850

851 Harvey, A.L., 2001. Twenty years of dendrotoxins. Toxicon 39, 15-26.

852

853 Izidoro, L.F.M., Ribeiro, M.C., Souza, G.R., Sant’Ana, C.D., Hamaguchi, A., Homsi-

854 Brandeburgo, M.I., Goulart, L.R., Beleboni, R.O., Nomizo, A., Sampaio, S.V., 2006.

855 Biochemical and functional characterization of an l-amino acid oxidase isolated from

856 Bothrops pirajai snake venom. Bioorg. Med. Chem. 14, 7034-7043.

857

858 Jia, L., Shimokawa, K., Bjarnason, J.B., Fox, J.W., 1996. Snake venom metalloproteinases:

859 structure, function and relationship to the ADAMs family of proteins. Toxicon 34, 1269-

860 1276.

861

862 Kardong, K., Weinstein, S., Smith, T., 2009. Reptile venom glands: form, function, and

863 future. In: Mackessy, S (Ed.), Handbook of venoms and toxins of reptiles, CRC Press, Boca

864 Raton, Florida, pp. 65-91.

865

866 Kasturiratne, A., Wickremasinghe, A.R., de Silva, N., Gunawardena, N.K., Pathmeswaran,

867 A., Premaratna, R., Savioli, L., Lalloo, D.G., de Silva, H.J., 2008. The global burden of

868 snakebite: a literature analysis and modelling based on regional estimates of envenoming and

869 deaths. PLoS medicine 5, e218. 34

870

871 Kemparaju, K., Girish, K., 2006. Snake venom hyaluronidase: a therapeutic target. Cell

872 Biochem. Funct. 24, 7-12.

873

874 Kini, R.M., Doley, R., 2010. Structure, function and evolution of three-finger toxins: mini

875 proteins with multiple targets. Toxicon 56, 855-867.

876

877 Kochva, E., 1987. The origin of snakes and evolution of the venom apparatus. Toxicon 25,

878 65-106.

879

880 Komori, Y., Nikai, T., 1998. Chemistry and Biochemistry of Kallikrein-Like Enzyn from

881 Snake Venoms. Toxin Reviews 17, 261-277.

882

883 Komori, Y., Nikai, T., Sugihara, H., 1988. Biochemical and physiological studies on a

884 kallikrein-like enzyme from the venom of Crotalus viridis viridis (Prairie rattlesnake).

885 Biochim. Biophys. Acta 967, 92-102.

886

887 Kostiza, T., Meier, J., 1996. Nerve growth factors from snake venoms: chemical properties,

888 mode of action and biological significance. Toxicon 34, 787-806.

889

890 Kulkeaw, K., Chaicumpa, W., Sakolvaree, Y., Tongtawe, P., Tapchaisri, P., 2007. Proteome

891 and immunome of the venom of the Thai cobra, Naja kaouthia. Toxicon 49, 1026-1041.

892

893 Kusumi, K., Kulathinal, R.J., Abzhanov, A., Boissinot, S., Crawford, N.G., Faircloth, B.C.,

894 Glenn, T.C., Janes, D.E., Losos, J.B., Menke, D.B., Poe, S., Sanger, T.J., Schneider, C.J.,

895 Stapley, J., Wade, J., Wilson-Rawls, J., 2011. Developing a community-based genetic

896 nomenclature for anole lizards. BMC Genomics 12, 554-2164-12-554.

897

898 Kwong, P.D., McDonald, N.Q., Sigler, P.B., Hendrickson, W.A., 1995. Structure of β2-

899 bungarotoxin: potassium channel binding by Kunitz modules and targeted phospholipase

900 action. Structure 3, 1109-1119.

901

902 Larkin, M.A., Blackshields, G., Brown, N.P., Chenna, R., McGettigan, P.A., McWilliam, H.,

903 Valentin, F., Wallace, I.M., Wilm, A., Lopez, R., Thompson, J.D., Gibson, T.J., Higgins,

904 D.G., 2007. Clustal W and Clustal X version 2.0. Bioinformatics 23, 2947-2948.

905

906 Li, M., Fry, B.G., Kini, R.M., 2005. Putting the brakes on snake venom evolution: the unique

907 molecular evolutionary patterns of Aipysurus eydouxii (Marbled sea snake) phospholipase A2

908 toxins. Mol. Biol. Evol. 22, 934-941.

909

910 Li, B., Dewey, C.N., 2011. RSEM: accurate transcript quantification from RNA-Seq data

911 with or without a reference genome. BMC Bioinformatics 12, 323-2105-12-323.

912

913 Lomonte, B., Escolano, J., Fernández, J., Sanz, L., Angulo, Y., Gutiérrez, J.M., Calvete, J.J.,

914 2008. Snake venomics and antivenomics of the arboreal neotropical pitvipers Bothriechis

915 lateralis and Bothriechis schlegelii. Journal of proteome research 7, 2445-2457.

916

917 Lynch, M., Conery, J.S., 2003. The evolutionary demography of duplicate genes. J. Struct.

918 Funct. Genomics 3, 35-44.

919

920 Lynch, V.J., 2007. Inventing an arsenal: adaptive evolution and neofunctionalization of snake

921 venom phospholipase A2 genes. BMC evolutionary biology 7, 2.

922

923 Lynch, M., Conery, J.S., 2000. The evolutionary fate and consequences of duplicate genes.

924 Science 290, 1151-1155.

925

926 Lynch, M., Force, A., 2000. The probability of duplicate gene preservation by

927 subfunctionalization. Genetics 154, 459-473.

928

929 Mackessy, S.P., 2002. Biochemistry and pharmacology of colubrid snake venoms. Toxin

930 Rev. 21, 43-83.

931

932 Margres, M.J., Aronow, K., Loyacano, J., Rokyta, D.R., 2013. The venom-gland

933 transcriptome of the Eastern coral snake (Micrurus fulvius) reveals high venom complexity in

934 the intragenomic evolution of venoms. BMC Genomics 14, 1-18.

935

936 Morita, T., 2005. Structures and functions of snake venom CLPs (C-type lectin-like proteins)

937 with anticoagulant-, procoagulant-, and platelet-modulating activities. Toxicon 45, 1099-

938 1114.

939

940 Nair, D., Fry, B., Alewood, P., Kumar, P., Kini, R., 2007. Antimicrobial activity of

941 omwaprin, a new member of the waprin family of snake venom proteins. Biochem. J. 402,

942 93-104.

943

944 Ogawa, T., Chijiwa, T., Oda-Ueda, N., Ohno, M., 2005. Molecular diversity and accelerated

945 evolution of C-type lectin-like proteins from snake venom. Toxicon 45, 1-14. 37

946

947 Oguiura, N., Boni-Mitake, M., Rádis-Baptista, G., 2005. New view on crotamine, a small

948 basic polypeptide myotoxin from South American rattlesnake venom. Toxicon 46, 363-370.

949

950 OmPraba, G., Chapeaurouge, A., Doley, R., Devi, K.R., Padmanaban, P., Venkatraman, C.,

951 Velmurugan, D., Lin, Q., Kini, R.M., 2010. Identification of a novel family of snake venom

952 proteins veficolins from Cerberus rynchops using a venom gland transcriptomics and

953 proteomics approach. J. Proteome Res. 9, 1882-1893.

954

955 Otto, S.P., Whitton, J., 2000. Polyploid incidence and evolution. Annu. Rev. Genet. 34, 401-

956 437.

957

958 Pahari, S., Mackessy, S.P., Kini, R.M., 2007. The venom gland transcriptome of the Desert

959 Massasauga Rattlesnake (Sistrurus catenatus edwardsii): towards an understanding of venom

960 composition among advanced snakes (Superfamily Colubroidea). BMC Mol. Biol. 8, 115.

961

962 Pirkle, H., 1998. Thrombin-like enzymes from snake venoms: an updated inventory. Thromb.

963 Haemost. 79, 675-683.

964

965 Pogrel, M.A., Low, M.A., Stern, R., 2003. Hyaluronan (hyaluronic acid) and its regulation in

966 human saliva by hyaluronidase and its inhibitors. J. Oral Sci. 45, 85-92.

967

968 Pung, Y.F., Kumar, S.V., Rajagopalan, N., Fry, B.G., Kumar, P.P., Kini, R.M., 2006. Ohanin,

969 a novel protein from king cobra venom: its cDNA and genomic organization. Gene 371, 246-

970 256.

971 38

972 Rádis-Baptista, G., Kubo, T., Oguiura, N., Svartman, M., Almeida, T., Batistic, R.F.,

973 Oliveira, E.B., Vianna-Morgante, ÂM., Yamane, T., 2003. Structure and chromosomal

974 localization of the gene for crotamine, a toxin from the South American rattlesnake, Crotalus

975 durissus terrificus. Toxicon 42, 747-752.

976

977 Richards, R., St Pierre, L., Trabi, M., Johnson, L.A., de Jersey, J., Masci, P.P., Lavin, M.F.,

978 2011. Cloning and characterisation of novel cystatins from elapid snake venom glands.

979 Biochimie 93, 659-668.

980

981 Ritonja, A., Evans, H.J., Machleidt, W., Barrett, A.J., 1987. Amino acid sequence of a

982 cystatin from venom of the African puff adder (Bitis arietans). Biochem. J. 246, 799-802.

983

984 Rokyta, D.R., Lemmon, A.R., Margres, M.J., Aronow, K., 2012. The venom-gland

985 transcriptome of the eastern diamondback rattlesnake (Crotalus adamanteus). BMC

986 Genomics 13, 312.

987

988 Rokyta, D.R., Wray, K.P., Lemmon, A.R., Lemmon, E.M., Caudle, S.B., 2011. A high-

989 throughput venom-gland transcriptome for the Eastern Diamondback Rattlesnake (Crotalus

990 adamanteus) and evidence for pervasive positive selection across toxin classes. Toxicon 57,

991 657-671.

992

993 Sanz, L., Escolano, J., Ferretti, M., Biscoglio, M.J., Rivera, E., Crescenti, E.J., Angulo, Y.,

994 Lomonte, B., Gutiérrez, J.M., Calvete, J.J., 2008. Snake venomics of the South and Central

995 American Bushmasters. Comparison of the toxin composition of Lachesis muta gathered

996 from proteomic versus transcriptomic analysis. J. Proteomics 71, 46-60.

997 39

998 Schwartz, T.S., Bronikowski, A.M., 2013. Dissecting molecular stress networks: identifying

999 nodes of divergence between life‐history phenotypes. Mol. Ecol. 22, 739-756.

1000

1001 Schwartz, T.S., Tae, H., Yang, Y., Mockaitis, K., Van Hemert, J.L., Proulx, S.R., Choi, J.,

1002 Bronikowski, A.M., 2010. A garter snake transcriptome: pyrosequencing, de novo assembly,

1003 and sex-specific differences. BMC Genomics 11, 694.

1004

1005 Serrano, S.M., Maroun, R.C., 2005. Snake venom serine proteinases: sequence homology vs.

1006 substrate specificity, a paradox to be solved. Toxicon 45, 1115-1132.

1007

1008 Siang, A.S., Doley, R., Vonk, F.J., Kini, R.M., 2010. Transcriptomic analysis of the venom

1009 gland of the red-headed krait (Bungarus flaviceps) using expressed sequence tags. BMC Mol.

1010 Biol. 11, 24-2199-11-24.

1011

1012 Strydom, D., 1973. Snake venom toxins: The evolution of some of the toxins found in snake

1013 venoms. Syst. Biol. 22, 596-608.

1014

1015 Suhr, S., Kim, D., 1996. Identification of the snake venom substance that induces apoptosis.

1016 Biochem. Biophys. Res. Commun. 224, 134-139.

1017

1018 Sunagar, K., Fry, B.G., Jackson, T.N., Casewell, N.R., Undheim, E.A., Vidal, N., Ali, S.A.,

1019 King, G.F., Vasudevan, K., Vasconcelos, V., 2013. Molecular Evolution of Vertebrate

1020 Neurotrophins: Co-Option of the Highly Conserved Nerve Growth Factor Gene into the

1021 Advanced Snake Venom Arsenal. PloS one 8, e81827.

1022

1023 Teshima, K.M., Innan, H., 2008. Neofunctionalization of duplicated genes under the pressure

1024 of gene conversion. Genetics 178, 1385-1398.

1025

1026 Torres, A.M., Wong, H.Y., Desai, M., Moochhala, S., Kuchel, P.W., Kini, R.M., 2003.

1027 Identification of a Novel Family of Proteins in Snake Venoms. Purification and structural

1028 characterization of nawaprin from Naja nigricollis snake venom. J. Biol. Chem. 278, 40097-

1029 40104.

1030

1031 Tu, A.T., Hendon, R.R., 1983. Characterization of lizard venom hyaluronidase and evidence

1032 for its action as a spreading factor. Comp. Biochem. Physiol. B. Comp. Biochem. 76, 377-

1033 383.

1034

1035 Tzika, A.C., Helaers, R., Schramm, G., Milinkovitch, M.C., 2011. Reptilian-transcriptome

1036 v1. 0, a glimpse in the brain transcriptome of five divergent Sauropsida lineages and the

1037 phylogenetic position of turtles. EvoDevo 2, 1-18.

1038

1039 Vidal, N., Hedges, S.B., 2005. The phylogeny of squamate reptiles (lizards, snakes, and

1040 amphisbaenians) inferred from nine nuclear protein-coding genes. Comptes rendus biologies

1041 328, 1000-1008.

1042

1043 Vikrant, S., Verma, B.S., 2013. Monitor lizard bite-induced acute kidney injury–a case

1044 report. Ren. Fail. 36, 444-446.

1045

1046 Vonk, F.J., Jackson, K., Doley, R., Madaras, F., Mirtschin, P.J., Vidal, N., 2011. Snake

1047 venom: From fieldwork to the clinic. Bioessays 33, 269-279.

1048 41

1049 Vonk, F.J., Casewell, N.R., Henkel, C.V., Heimberg, A.M., Jansen, H.J., McCleary, R.J.,

1050 Kerkkamp, H.M., Vos, R.A., Guerreiro, I., Calvete, J.J., Wuster, W., Woods, A.E., Logan,

1051 J.M., Harrison, R.A., Castoe, T.A., de Koning, A.P., Pollock, D.D., Yandell, M., Calderon,

1052 D., Renjifo, C., Currier, R.B., Salgado, D., Pla, D., Sanz, L., Hyder, A.S., Ribeiro, J.M.,

1053 Arntzen, J.W., van den Thillart, G.E., Boetzer, M., Pirovano, W., Dirks, R.P., Spaink, H.P.,

1054 Duboule, D., McGlinn, E., Kini, R.M., Richardson, M.K., 2013. The king cobra genome

1055 reveals dynamic gene evolution and adaptation in the snake venom system. Proc. Natl. Acad.

1056 Sci. U. S. A. 110, 20651-20656.

1057

1058 Wagstaff, S.C., Harrison, R.A., 2006. Venom gland EST analysis of the saw-scaled viper,

1059 Echis ocellatus, reveals novel α9 β1 integrin-binding motifs in venom metalloproteinases and

1060 a new group of putative toxins, renin-like aspartic proteases. Gene 377, 21-32.

1061

1062 Wagstaff, S.C., Laing, G.D., Theakston, R.D.G., Papaspyridis, C., Harrison, R.A., 2006.

1063 Bioinformatics and multiepitope DNA immunization to design rational snake antivenom.

1064 PLoS Med. 3, e184.

1065

1066 Wagstaff, S.C., Sanz, L., Juárez, P., Harrison, R.A., Calvete, J.J., 2009. Combined snake

1067 venomics and venom gland transcriptomic analysis of the ocellated carpet viper, Echis

1068 ocellatus. J. Proteomics 71, 609-623.

1069

1070 Walsh, B., 2003. Population-genetic models of the fates of duplicate genes. Genetica 118,

1071 279-294.

1072

1073 Warner, T.G., Dambach, L.M., Shin, J.H., O'Brien, J.S., 1981. Purification of the lysosomal

1074 acid lipase from human liver and its role in lysosomal lipid hydrolysis. J. Biol. Chem. 256,

1075 2952-2957.

1076

1077 Whitton, J.L., Sheng, N., Oldstone, M.B., McKee, T.A., 1993. A "string-of-beads" vaccine,

1078 comprising linked minigenes, confers protection from lethal-dose virus challenge. J. Virol.

1079 67, 348-352.

1080

1081 Williford, A., Demuth, J.P., 2012. Gene expression levels are correlated with synonymous

1082 codon usage, amino acid composition, and gene architecture in the red flour beetle, Tribolium

1083 castaneum. Mol. Biol. Evol. 29, 3755-3766.

1084

1085 Wong, E.S., Belov, K., 2012. Venom evolution through gene duplications. Gene 496, 1-7.

1086

1087 Yamazaki, Y., Hyodo, F., Morita, T., 2003a. Wide distribution of cysteine-rich secretory

1088 proteins in snake venoms: isolation and cloning of novel snake venom cysteine-rich secretory

1089 proteins. Arch. Biochem. Biophys. 412, 133-141.

1090

1091 Yamazaki, Y., Morita, T., 2004. Structure and function of snake venom cysteine-rich

1092 secretory proteins. Toxicon 44, 227-231.

1093

1094 Yamazaki, Y., Takani, K., Atoda, H., Morita, T., 2003b. Snake venom vascular endothelial

1095 growth factors (VEGFs) exhibit potent activity through their specific recognition of KDR

1096 (VEGF receptor 2). J. Biol. Chem. 278, 51985-51988.

1097

1098 Župunski, V., Kordiš, D., Gubenšek, F., 2003. Adaptive evolution in the snake venom

1099 Kunitz/BPTI protein family. FEBS Lett. 547, 131-136.

1100

1101

1102

1103

1104

1105

1106

1107

1108

1109

1110

1111

1112

1113

1114 Tables

1115

1116 Table 1. Predicted venom composition of the painted saw-scaled viper, Echis coloratus

1117

Gene family Number of genes

SVMP 13

C-type lectin 8

Serine protease 6

PLA2 3

CRISP 1

L-amino acid oxidase 1

VEGF 1

Crotamine 1

Total 8 34

1118

1119

1120

1121

1122

1123

1124

1125

1126

1127

1128

1129 45

1130 Table 2. Predicted numbers of venom toxins and venom toxin families from proteomic studies

1131 of snake venom accord well with our transcriptome results.

1132

Species Number of Number of toxin

toxins families

Bitis caudalis (Calvete et al., 2007a) 30 8

Bitis gabonica gabonica (Calvete et 35 12 al., 2006)

Bitis gabonica rhinoceros (Calvete 33 11 et al., 2007a)

Bitis nasicornis (Calvete et al., 28 9 2007a)

Bothriechis schlegelii (Lomonte et ? 7 al., 2008)

Cerastes cerastes (Fahmi et al., 25-30 6 2012)

Crotalus atrox (Calvete et al., 2009) ~24 ~9

Echis ocellatus (Wagstaff et al., 35 8 2009)

Lachesis muta (Sanz et al., 2008) 24-26 8

Naja kaouthia (Kulkeaw et al., 61 12 2007)

Ophiophagus hannah (Vonk et al., ? 14 2013)

Vipera ammodytes (Georgieva et al., 38 9 2008)

1133

1134

1135

1136

1137

1138

1139

1140

1141

1142

1143

1144

1145

1146

1147 Table 3. Venom gene nomenclature. Lack of a formal set of nomenclatural rules for venom

1148 toxins has led to an explosion of different gene names and may have contributed to the

1149 overestimation of reptile venom diversity.

Gene/gene family Alternative name and accession number

3 Finger toxin (3Ftx) Denmotoxin [Q06ZW0]

Candoxin [AY142323]

CRISP Piscivorin [AAO62994]

Catrin [AAO62995]

Ablomin [AAM45664]

Tigrin [Q8JGT9]

Kaouthin [ACH73167, ACH73168]

Natrin-1 [Q7T1K6]

CRVP [Q8UW25, Q8UW11]

Pseudechetoxin [Q8AVA4]

Pseudechin [Q8AVA3]

Serotriflin [P0CB15]

Latisemin [Q8JI38]

Ophanin [AAO62996]

Opharin [ACN93671]

Bc-CRP [ACE73577, ACE73578]

Ficolin Veficolin [ADK46899]

Ryncolin [D8VNS7-9, D8VNT0]

Serine proteases Acubin [CAB46431]

Gyroxin [B0FXM3]

Ussurase [AAL48222]

Serpentokallikrein [AAG27254]

Salmobin [AAC61838]

Batroxobin [AAA48553]

Nikobin [CBW30778]

Gloshedobin [POC5B4]

Gussurobin [Q8UVX1]

Pallabin [CAA04612]

Pallase [AAC34898]

Snake venom metalloproteinase Stejnihagin-B [ABA40759]

(SVMP) Bothropasin [AAC61986]

Atrase B [ADG02948]

Mocarhagin 1 [AAM51550]

Scutatease-1 [ABQ01138]

Austrelease-1 [ABQ01134]

Vascular endothelial growth factor Barietin [ACN22038]

(VEGF) Cratrin [ACN22040]

Apiscin [ACN22039]

Vammin [ACN22045]

Vespryn Ohanin [AAR07992]

Thaicobrin [P82885]

Waprin Nawaprin [P60589]

Porwaprin [B5L5N2]

Stewaprin [B5G6H3]

Veswaprin [B5L5P5]

Notewaprin [B5G6H5]

Carwaprin [B5L5P0]

1150

1151

1152 Figure legends

1153

1154 Figure 1. Relationships of key vertebrate lineages and the placement of species

1155 discussed in this paper. A monophyletic clade of reptiles (which includes birds) is

1156 boxed and the Toxicofera (Fry et al., 2013) are shaded. Modified taxon names

1157 have been used for simplicity. Due to the lack of taxonomic resolution within the Colubridae,

1158 we have placed the term colubrids in inverted commas.

1159

1160 Figure 2. Tissue distribution of proposed venom toxin transcripts. The majority of

1161 transcripts proposed to encode Toxicoferan venom proteins are expressed in multiple body

1162 tissues. Transcript order follows descriptions in the main text and those transcripts found in

1163 the assembled transcriptomes but which are assigned transcript abundance of <1 FPKM are

1164 shaded orange. VG, venom gland; SAL, salivary gland; SCG, scent gland; SK, skin.

1165

1166 Figure 3. Maximum likelihood tree of complement c3 (“cobra venom factor”) sequences.

1167 Whilst most sequences likely represent housekeeping or maintenance genes, a gene

1168 duplication event in the elapid lineage (marked with *) may have produced a venom-specific

1169 paralog. An additional duplication (marked with +) may have taken place in Austrelaps

1170 superbus, although both paralogs appear to be expressed in both liver and venom gland.

1171 Geographic separation in king cobras (Ophiophagus hannah) from Indonesia and China is

1172 reflected in observed sequence variation. Numbers above branches are Bootstrap values for

1173 500 replicates. Tissue distribution of transcripts is indicated using the following

1174 abbreviations: VG, venom gland; SK, skin; SCG, scent gland, AG, accessory gland; VMNO,

1175 vomeronasal organ and those genes found to be expressed in one or more body tissues are

1176 shaded blue.

1177 51

1178 Figure 4. Maximum likelihood tree of dipeptidylpeptidase 3 (dpp3) and

1179 dipeptidylpeptidase 4 (dpp4) sequences. Transcripts encoding dpp3 and dpp4 are found in a

1180 wide variety of body tissues, and likely represent housekeeping genes. Numbers above

1181 branches are Bootstrap values for 500 replicates. Tissue distribution of transcripts is indicated

1182 using the following abbreviations: VG, venom gland; SK, skin; SCG, scent gland, AG,

1183 accessory gland; VMNO, vomeronasal organ and those genes found to be expressed in one or

1184 more body tissues are shaded blue.

1185

1186 Figure 5. Maximum likelihood tree of phospholipase b (plb) sequences. Transcripts

1187 encoding plb are found in a wide variety of body tissues, and likely represent housekeeping

1188 genes. Numbers above branches are Bootstrap values for 500 replicates. Tissue distribution of

1189 transcripts is indicated using the following abbreviations: VG, venom gland; SK, skin; SCG,

1190 scent gland, AG, accessory gland; VMNO, vomeronasal organ and those genes found to be

1191 expressed in one or more body tissues are shaded blue.

1192

1193 Figure 6. Maximum likelihood tree of renin-like sequences. Renin-like genes are

1194 expressed in a diversity of body tissues. The recently published Boa constrictor “RAP-Boa-

1195 1” sequence is clearly a cathepsin d gene and is therefore not orthologous to the Echis

1196 ocellatus renin sequence as has been claimed (Fry et al., 2013). Numbers above branches are

1197 Bootstrap values for 500 replicates. Tissue distribution of transcripts is indicated using the

1198 following abbreviations: VG, venom gland; SK, skin; SCG, scent gland and those genes

1199 found to be expressed in one or more body tissues are shaded blue.

1200

1201 Figure 7. Graph of transcript abundance values of proposed venom transcripts in the

1202 Echis coloratus venom gland. The majority of Toxicoferan transcripts are expressed at

1203 extremely low level, with the most highly expressed genes falling into only four gene 52

1204 families (C-type lectins, Group IIA phospholipase A2, serine proteases and snake venom

1205 metalloproteinases). FPKM = Fragments Per Kilobase of exon per Million fragments

1206 mapped.